Session:<\/strong> Networking Correspondences<\/p>\n\n\n\n
Archives have become essential components in the \u2018laboratories\u2019 of those doing historical network research, rendering us with the metadata and digitised collections for computational experiments. However, we have not always taken the time to understand how these sources have come into being (Walsham 2016). Historical correspondence archives, in particular, have been a rich source of relational data on which to base historical network research (for example B\u0142och, Vasques Filho, and Bojanowski 2020; Ahnert and Ahnert 2019). While authors often reference the fragmentary nature of their archives as a way of encouraging caution regarding their interpretations, the size, shape and gaps in these sources are rarely discussed in the quantitative terms now often used with other sources, such as web archives (Br\u00fcgger 2018, 74). <\/p>\n\n\n\n
Outside of the digital humanities, the \u2018archival turn\u2019\u2014an approach which considers an archive as not\u00a0 a neutral collection but rather a text in its own right, one which is what Penman (2016) calls \u2018the\u00a0 product of a multiplicity of interventions\u2019\u2014has become an important area of scholarship. This paper\u00a0 argues that if an archive can be considered a text, then it can itself be the subject of the \u2018macroanalysis\u2019\u00a0 long espoused by digital humanities practitioners (Jockers 2013). Rather than using correspondence\u00a0 archives to extract and analyse the networks found within them, this paper will outline how network\u00a0 methods can be leveraged to write or update histories of archives themselves. Using network analysis,\u00a0 it will examine two seemingly-different sets of sources, assembled at different times, and give some\u00a0 specific examples of how network analysis can aid the understanding of the historical contingencies\u00a0 and collection practices which caused these archives to have the structure and shape they do.\u00a0\u00a0<\/p>\n\n\n\n
The Networking Archives <\/em>project, from which this work arises, is analysing metadata from three\u00a0 separate sets of Early Modern (1500-1800) correspondence, which have been cleaned and reconciled\u00a0 extensively: the Tudor and Stuart State Papers, and Early Modern Letters Online (EMLO). Together,\u00a0 they form a dataset of c. 450,000 letters, and can be represented as a graph of over 60,000 nodes and\u00a0 100,000 edges, with substantial overlap between them. It is perhaps the largest historical\u00a0 correspondence dataset so far assembled in one place, but the constituent parts are ultimately from\u00a0 reassembled, partial sources. One, EMLO, is the product of a twenty-first century effort to collect\u00a0 and combine a wide range of epistolary sources relating to the Republic of Letters, whereas the other\u00a0 two are the \u2018official\u2019 archives of the working papers of the English Secretaries of State. Despite this\u00a0 official status of the latter, the history of the State Papers is complex: the lines between public and\u00a0 private documents were blurry, and many papers which should have been part of the archive were\u00a0 lost or kept in private hands and not reunited with the rest for several centuries\u2014the Conway Papers\u00a0 being one example (Smith 2014). The result is that despite these very different origins, when analysed\u00a0 as a network, the archives share many similar properties\u2014for example, the degree distribution of\u00a0 each looks remarkably similar and is roughly \u2018scale-free\u2019 (figure 1) .\u00a0\u00a0<\/p>\n\n\n\n The paper describes how some key network analysis methods and concepts\u2014degree, strength, clustering coefficient, connected components and the rich club\u2014might be used to understand the shape and structure of archives at an aggregate level. It will show how connected and disconnected components, for example, can highlight caches of confiscated family papers and bundles of correspondence seized from captured ships. It will describe how some basic network metrics\u2014degree and weighted degree\u2014 when considered on a global scale, can serve as a useful comparative measure to understand the \u2018type\u2019 of archive in question, particularly when the direction of the links are considered. An analysis of the clustering coefficient of the SPO and EMLO data will highlight the extent to which, on the whole, these are tightly-linked sources, and the extraction of the rich club and rich club coefficient will be used to suggest the extent to which each is an archive of \u2018elites.\u2019 The paper will also\u2014briefly\u2014discuss the use of network analysis to measure the gaps present in an archive, and outline how this might help us to estimate portions of missing data and assess its impact on quantitative network results. <\/p>\n\n\n\n Applying network tools to the study of entire archives has revealed that even seemingly-homogeneous sources such as the State Papers should be considered as partial, reassembled and based on contingency. It follows that rather than think about each as individual silos, they should be viewed as an overlapping set, a fact which allows comparisons between them on an aggregate level. The paper will argue, moreover, that network tools allow us to ignore or supplement the often-artificial boundaries of archives\u2014the catalogues, series or folios in which they have been collected and ordered\u2014and reshuffle the sources in any way we see fit. It will consider the opportunities and pitfalls of this way of thinking about historical archives, and the impact it may have in the field of historical network research. <\/p>\n\n\n\n Works Cited <\/strong><\/p>\n\n\n\n Ahnert, Ruth, and Sebastian E Ahnert. 2019. \u201cMetadata, Surveillance and the Tudor State.\u201d History Workshop Journal <\/em>87: 27\u201351. https:\/\/doi.org\/10.1093\/hwj\/dby033<\/a>.\u00a0<\/p>\n\n\n\n Br\u00fcgger, Niels. 2018. The Archived Web: Doing History in the Digital Age<\/em>. Cambridge, Massachusetts: The MIT Press. <\/p>\n\n\n\n B\u0142och, Agata, Demival Vasques Filho, and Micha\u0142 Bojanowski. 2020. \u201cNetworks from Archives: Recon structing Networks of Official Correspondence in the Early Modern Portuguese Empire.\u201d Social Networks<\/em>, September. https:\/\/doi.org\/10.1016\/j.socnet.2020.08.008<\/a>.\u00a0<\/p>\n\n\n\n Jockers, Matthew L. 2013. Macroanalysis: Digital Methods and Literary History<\/em>. University of Illinois Press. https:\/\/www.jstor.org\/stable\/10.5406\/j.ctt2jcc3m<\/a>.\u00a0<\/p>\n\n\n\n Penman, Leigh T. I. 2016. \u201cOmnium Exposita Rapin\u00e6: The Afterlives of the Papers of Samuel Hartlib.\u201d Book History <\/em>19 (1): 1\u201365. https:\/\/doi.org\/10.1353\/bh.2016.0000<\/a>.\u00a0<\/p>\n\n\n\n Smith, Daniel Starza. 2014. The Curious History of the Conway Papers<\/em>. Oxford University Press. https:\/\/doi.org\/10.1093\/acprof:oso\/9780199679133.003.0007<\/a>.\u00a0<\/p>\n\n\n\n Walsham, Alexandra. 2016. \u201cThe Social History of the Archive: Record-Keeping in Early Modern Europe.\u201d Past & Present <\/em>230 (November): 9\u201348. https:\/\/doi.org\/10.1093\/pastj\/gtw033<\/a>. <\/p>\n","protected":false},"excerpt":{"rendered":" Yann Ciar\u00e1n Ryan Time and Place: Thursday, 01.07., 11:35\u201311:55, Room 2Session: Networking Correspondences Archives have become essential components in the \u2018laboratories\u2019 of those doing historical network research, rendering us with the metadata and digitised collections for computational experiments. However, we have not always taken the time to understand how these sources have come into being (Walsham 2016). Historical correspondence archives,<\/p>\n![\"\"](\"http:\/\/hnr2021.historicalnetworkresearch.org\/wp-content\/uploads\/2021\/06\/HNRResHist2021_paper_39fig1.png\")
<\/a>